MiniMax-M2.5 is a large language model built for real-world productivity and trained across many complex digital work environments. It extends the coding strengths of M2.1 to office tasks: creating and editing Word, Excel, and PowerPoint files, switching context between software tools, and working in mixed agent and human teams. It scores 80.2% on SWE-Bench Verified, 51.3% on Multi-SWE-Bench, and 76.3% on BrowseComp. It is also more token-efficient than earlier generations, having been trained to plan its actions and outputs.
$0.135market price $0.27
$0.0135market price $0.027
$0.475market price $0.95
Throughput is how fast the model writes (tokens per second, higher is better). Time to first token is how long before the answer starts (lower is better).
Share of requests served without a server error. Client errors (4xx) do not count.
Works with the OpenAI SDK pointed at our base URL. Every parameter is in the API reference.
Tokens processed per day on NagaAI.
MiniMax-M2.5 is a large language model built for real-world productivity and trained across many complex digital work environments.
On NagaAI, MiniMax M2.5 costs $0.135 per 1M input tokens and $0.475 per 1M output tokens. You pay only for what you use, with no subscription.
MiniMax M2.5 reads up to 204.8k tokens in one request, counting the prompt and the answer together.
Yes. MiniMax M2.5 can call your functions as tools and can answer in a JSON schema you give it.
It takes text as input and returns text.
Create an API key, copy the model ID from the top of this page and send requests with the OpenAI or Anthropic SDK pointed at NagaAI. The API section on this page has a ready example.
MiniMax M3.1 FlashMinimaxMiniMax M3MinimaxMiniMax M2.7MinimaxMiniMax M2.1Minimax$0.15 / $0.60 per 1MMiniMax M2Minimax$0.128 / $0.51 per 1MMiniMax M1MinimaxNot enough throughput data
Not enough TTFT data